Project: Vehicle Detection¶
In [1]:
import random
import matplotlib.image as mpimg
import matplotlib.pyplot as plt
import numpy as np
import cv2
import glob
import time
from sklearn.svm import LinearSVC
from sklearn.preprocessing import StandardScaler
from skimage.feature import hog
from sklearn.model_selection import train_test_split
from scipy.ndimage.measurements import label
In [2]:
# Divide up into cars and notcars
cars = glob.glob('dataset/vehicles/**/*.png')
notcars = glob.glob('dataset/non-vehicles/**/*.png')
# generate 10 random car and non-car images
car_inds = random.sample(range(0, len(cars)-1), 10)
notcar_inds = random.sample(range(0, len(notcars)-1), 10)
In [3]:
# function to read image in rgb format, video frame is by-default in rgb format
def read_image(file):
# reads image in BGR format
image = cv2.imread(file)
# convert to RGB and return
return cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
In [4]:
def convert_color(image, cspace):
# apply color conversion if other than 'RGB'
if cspace != 'RGB':
if cspace == 'HSV':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2HSV)
elif cspace == 'LUV':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2LUV)
elif cspace == 'HLS':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2HLS)
elif cspace == 'YUV':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2YUV)
elif cspace == 'YCrCb':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2YCrCb)
else: feature_image = np.copy(image)
return feature_image
In [5]:
# display dataset samples
print('Total car images : ', len(cars))
print('Total non-car images : ', len(notcars))
# display 10 random car images and non-car images
def show_images(img_list, inds):
fg, ax = plt.subplots(1, 10, figsize=(20, 5))
ax = ax.ravel()
for i in range(10):
ax[i].axis('off')
ax[i].imshow(read_image(img_list[inds[i]]))
ax[i].set_title(inds[i])
show_images(cars, car_inds)
show_images(notcars, notcar_inds)
In [6]:
# Define a function to return HOG features and visualization
def get_hog_features(img, orient, pix_per_cell, cell_per_block, vis=False, feature_vec=True):
# Call with two outputs if vis==True
if vis == True:
features, hog_image = hog(img, orientations=orient, pixels_per_cell=(pix_per_cell, pix_per_cell),
cells_per_block=(cell_per_block, cell_per_block), block_norm= 'L2-Hys',
transform_sqrt=True,
visualize=vis, feature_vector=feature_vec)
return features, hog_image
# Otherwise call with one output
else:
features = hog(img, orientations=orient, pixels_per_cell=(pix_per_cell, pix_per_cell),
cells_per_block=(cell_per_block, cell_per_block), block_norm= 'L2-Hys',
transform_sqrt=True,
visualize=vis, feature_vector=feature_vec)
return features
In [7]:
# try HOG for different color spaces
def show_color_hog(file):
fg, ax = plt.subplots(3, 8, figsize=(20, 8))
ax = ax.ravel()
i = 0
img = read_image(file)
# color spaces to experiment
cspace = ["RGB", "LUV", "HLS", "YUV", "HSV", "YCrCb"]
for color_space in cspace:
ax[i].axis('off')
ax[i].imshow(img)
ax[i].set_title(color_space)
img_out = convert_color(img, color_space)
for j in range(3):
img_ch = img_out[:,:,j]
hog_features, hog_image = get_hog_features(img_ch, 9, 8, 2, True)
ax[i+1+j].axis('off')
ax[i+1+j].imshow(hog_image, cmap='gray')
ax[i+1+j].set_title(j)
i = i + 4
show_color_hog(cars[car_inds[0]])
show_color_hog(notcars[notcar_inds[0]])
In [8]:
# try HOG with different parameters
def show_hog_params(file):
fg, ax = plt.subplots(3, 8, figsize=(20, 8))
ax = ax.ravel()
i = 0
img = read_image(file)
hog_vals = [ [9, 8, 2], [11, 8, 2], [11, 16, 2] ]
cspace = [ 'YUV', 'YCrCb' ]
for hog_val in hog_vals:
for color in cspace:
ax[i].axis('off')
ax[i].imshow(img)
ax[i].set_title(color)
img_out = convert_color(img, color)
for j in range(3):
img_ch = img_out[:,:,j]
hog_features, hog_image = get_hog_features(img_ch, hog_val[0], hog_val[1], hog_val[2], True)
ax[i+1+j].axis('off')
ax[i+1+j].imshow(hog_image, cmap='gray')
ax[i+1+j].set_title('ch'+str(j)+' : '+str(hog_val))
i = i + 4
show_hog_params(cars[car_inds[0]])
In [9]:
# combine HOG features of all channels
def extract_hog_features(image, cspace, orient, pix_per_cell, cell_per_block, hog_channel):
# apply color conversion if other than 'RGB'
feature_image = convert_color(image, cspace)
# Call get_hog_features() with vis=False, feature_vec=True
if hog_channel == 'ALL':
hog_features = []
for channel in range(feature_image.shape[2]):
hog_features.append(get_hog_features(feature_image[:,:,channel], orient,
pix_per_cell, cell_per_block, vis=False, feature_vec=True))
hog_features = np.ravel(hog_features)
else:
hog_features = get_hog_features(feature_image[:,:,hog_channel], orient,
pix_per_cell, cell_per_block, vis=False, feature_vec=True)
return hog_features
In [10]:
# Define a function to extract features from a list of images
def extract_features(imgs, cspace, orient, pix_per_cell, cell_per_block, hog_channel):
# Create a list to append feature vectors to
features = []
# Iterate through the list of images
for file in imgs:
# Read in each one by one
image = read_image(file)
# extract hog features
hog_features = extract_hog_features(image, cspace, orient, pix_per_cell, cell_per_block, hog_channel)
# Append the new feature vector to the features list
features.append(hog_features)
# Return list of feature vectors
return features
In [11]:
# train the classifier
def train_clf(colorspace, hog_channel, orient, pix_per_cell, cell_per_block):
### print('Using:', colorspace, hog_channel, 'channel,',
### orient,'orientations,',
### pix_per_cell,'pixels per cell and',
### cell_per_block,'cells per block')
t1=time.time()
car_features = extract_features(cars, cspace=colorspace, orient=orient,
pix_per_cell=pix_per_cell, cell_per_block=cell_per_block,
hog_channel=hog_channel)
notcar_features = extract_features(notcars, cspace=colorspace, orient=orient,
pix_per_cell=pix_per_cell, cell_per_block=cell_per_block,
hog_channel=hog_channel)
t2 = time.time()
### print(round(t2-t1, 2), 'Seconds to extract HOG features...')
# Create an array stack of feature vectors
X = np.vstack((car_features, notcar_features)).astype(np.float64)
# Define the labels vector
y = np.hstack((np.ones(len(car_features)), np.zeros(len(notcar_features))))
# Split up data into randomized training and test sets
rand_state = np.random.randint(0, 100)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=rand_state)
# Fit a per-column scaler
X_scaler = StandardScaler().fit(X_train)
# Apply the scaler to X
X_train = X_scaler.transform(X_train)
X_test = X_scaler.transform(X_test)
### print('Train dataset len', len(X_train))
### print('Test dataset len', len(X_test))
### print('Feature vector length:', len(X_train[0]))
# Use a linear SVC
svc = LinearSVC()
# Check the training time for the SVC
t3=time.time()
svc.fit(X_train, y_train)
t4 = time.time()
### print(round(t4-t3, 2), 'Seconds to train SVC...')
# Check the score of the SVC
### print('Train Accuracy of SVC = ', round(svc.score(X_train, y_train), 4))
### print('Test Accuracy of SVC = ', round(svc.score(X_test, y_test), 4))
# Check the prediction time for a single sample
#t=time.time()
#n_predict = 10
#print('My SVC predicts: ', svc.predict(X_test[0:n_predict]))
#print('For these',n_predict, 'labels: ', y_test[0:n_predict])
#t2 = time.time()
#print(round(t2-t, 5), 'Seconds to predict', n_predict,'labels with SVC')
print(colorspace, hog_channel, orient, pix_per_cell, cell_per_block, round(t2-t1, 2),
len(X_train), len(X_test), len(X_train[0]), round(t4-t3, 2),
round(svc.score(X_train, y_train), 4), round(svc.score(X_test, y_test), 4))
return svc, X_scaler
In [12]:
print('colorspace', 'channel', 'orient', 'pix/cell', 'cell/block', 'T-extract',
'N-train', 'N-test', 'N-features', 'T-train', 'Train-%', 'Test-%')
#clf111 = train_clf('YUV', 0, 9, 8, 2)
#clf112 = train_clf('YUV', 1, 9, 8, 2)
#clf113 = train_clf('YUV', 2, 9, 8, 2)
#clf114, X_scaler114 = train_clf('YUV', 'ALL', 9, 8, 2)
#clf121 = train_clf('YUV', 0, 11, 8, 2)
#clf122 = train_clf('YUV', 1, 11, 8, 2)
#clf123 = train_clf('YUV', 2, 11, 8, 2)
#clf124, X_scaler124 = train_clf('YUV', 'ALL', 11, 8, 2)
#clf131 = train_clf('YUV', 0, 11, 16, 2)
#clf132 = train_clf('YUV', 1, 11, 16, 2)
#clf133 = train_clf('YUV', 2, 11, 16, 2)
#clf134, X_scaler134 = train_clf('YUV', 'ALL', 11, 16, 2)
#clf211 = train_clf('YCrCb', 0, 9, 8, 2)
#clf212 = train_clf('YCrCb', 1, 9, 8, 2)
#clf213 = train_clf('YCrCb', 2, 9, 8, 2)
#clf214, X_scaler214 = train_clf('YCrCb', 'ALL', 9, 8, 2)
#clf221 = train_clf('YCrCb', 0, 11, 8, 2)
#clf222 = train_clf('YCrCb', 1, 11, 8, 2)
#clf223 = train_clf('YCrCb', 2, 11, 8, 2)
#clf224, X_scaler224 = train_clf('YCrCb', 'ALL', 11, 8, 2)
#clf231 = train_clf('YCrCb', 0, 11, 16, 2)
#clf232 = train_clf('YCrCb', 1, 11, 16, 2)
#clf233 = train_clf('YCrCb', 2, 11, 16, 2)
clf234, X_scaler234 = train_clf('YCrCb', 'ALL', 11, 16, 2)
In [13]:
# sliding window search
def find_cars(img, cspace, ystart, ystop, scale, svc, X_scaler, orient, pix_per_cell, cell_per_block, disp=False):
# empty array to save rectangles
rects = []
# list of search windows, just for display
windows = []
img_tosearch = img[ystart:ystop,:,:]
ctrans_tosearch = convert_color(img_tosearch, cspace)
if scale != 1:
imshape = ctrans_tosearch.shape
ctrans_tosearch = cv2.resize(ctrans_tosearch, (np.int(imshape[1]/scale), np.int(imshape[0]/scale)))
ch1 = ctrans_tosearch[:,:,0]
ch2 = ctrans_tosearch[:,:,1]
ch3 = ctrans_tosearch[:,:,2]
# Define blocks and steps as above
nxblocks = (ch1.shape[1] // pix_per_cell) - cell_per_block + 1
nyblocks = (ch1.shape[0] // pix_per_cell) - cell_per_block + 1
nfeat_per_block = orient*cell_per_block**2
# 64 was the orginal sampling rate, with 8 cells and 8 pix per cell
window = 64
nblocks_per_window = (window // pix_per_cell) - cell_per_block + 1
cells_per_step = 2 # Instead of overlap, define how many cells to step
nxsteps = (nxblocks - nblocks_per_window) // cells_per_step + 1
nysteps = (nyblocks - nblocks_per_window) // cells_per_step + 1
# Compute individual channel HOG features for the entire image
hog1 = get_hog_features(ch1, orient, pix_per_cell, cell_per_block, feature_vec=False)
hog2 = get_hog_features(ch2, orient, pix_per_cell, cell_per_block, feature_vec=False)
hog3 = get_hog_features(ch3, orient, pix_per_cell, cell_per_block, feature_vec=False)
for xb in range(nxsteps):
for yb in range(nysteps):
ypos = yb*cells_per_step
xpos = xb*cells_per_step
# Extract HOG for this patch
hog_feat1 = hog1[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel()
hog_feat2 = hog2[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel()
hog_feat3 = hog3[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel()
hog_features = np.hstack((hog_feat1, hog_feat2, hog_feat3))
xleft = xpos*pix_per_cell
ytop = ypos*pix_per_cell
test_features = X_scaler.transform(np.hstack((hog_features)).reshape(1, -1))
test_prediction = svc.predict(test_features)
if disp == True:
xbox_left = np.int(xleft*scale)
ytop_draw = np.int(ytop*scale)
win_draw = np.int(window*scale)
windows.append([(xbox_left, ytop_draw+ystart),(xbox_left+win_draw,ytop_draw+win_draw+ystart)])
if test_prediction == 1:
xbox_left = np.int(xleft*scale)
ytop_draw = np.int(ytop*scale)
win_draw = np.int(window*scale)
rects.append([(xbox_left, ytop_draw+ystart),(xbox_left+win_draw,ytop_draw+win_draw+ystart)])
#cv2.rectangle(draw_img,(xbox_left, ytop_draw+ystart),(xbox_left+win_draw,ytop_draw+win_draw+ystart),(0,0,255),6)
if disp == True:
return rects, windows
return rects
In [14]:
def draw_boxes(img, bboxes, color=(0, 0, 255), thick=6):
# Make a copy of the image
imcopy = np.copy(img)
# Iterate through the bounding boxes
for bbox in bboxes:
# Draw a rectangle given bbox coordinates
cv2.rectangle(imcopy, bbox[0], bbox[1], color, thick)
# Return the image copy with boxes drawn
return imcopy
In [15]:
def show_windows(img, ystart, ystop, scale, orient, pix_per_cell, cell_per_block):
rects, windows = find_cars(img, 'YCrCb', ystart, ystop, scale, clf234, X_scaler234,
orient, pix_per_cell, cell_per_block, True)
rect_out = draw_boxes(img, rects, (0, 0, 255))
win_out = draw_boxes(img, windows, (0, 255, 0))
fg, ax = plt.subplots(1, 2, figsize=(40, 10))
ax = ax.ravel()
ax[0].imshow(win_out)
ax[0].set_title('sliding windows')
ax[1].imshow(rect_out)
ax[1].set_title('search output')
show_windows(read_image('test_images/test5.jpg'), 390, 662, 2, 11, 16, 2)
In [16]:
def add_heat(heatmap, bbox_list, weight = 1):
# Iterate through list of bboxes
for box in bbox_list:
# Add += 1 for all pixels inside each bbox
# Assuming each "box" takes the form ((x1, y1), (x2, y2))
heatmap[box[0][1]:box[1][1], box[0][0]:box[1][0]] += weight
# Return updated heatmap
return heatmap# Iterate through list of bboxes
In [17]:
def apply_threshold(heatmap, threshold):
# Zero out pixels below the threshold
heatmap[heatmap <= threshold] = 0
# Return thresholded map
return heatmap
In [18]:
def draw_labeled_bboxes(labels):
rects = []
# Iterate through all detected cars
for car_number in range(1, labels[1]+1):
# Find pixels with each car_number label value
nonzero = (labels[0] == car_number).nonzero()
# Identify x and y values of those pixels
nonzeroy = np.array(nonzero[0])
nonzerox = np.array(nonzero[1])
# Define a bounding box based on min/max x and y
bbox = ((np.min(nonzerox), np.min(nonzeroy)), (np.max(nonzerox), np.max(nonzeroy)))
# Draw the box on the image
#cv2.rectangle(img, bbox[0], bbox[1], (0,0,255), 6)
rects.append([bbox[0], bbox[1]])
# Return the image
return rects
In [19]:
# create heatmap and threshold
def find_heat(img, box_list, thresh):
heat = np.zeros_like(img[:,:,0]).astype(np.float)
heat = add_heat(heat,box_list)
heat = apply_threshold(heat,thresh)
heatmap = np.clip(heat, 0, 255)
labels = label(heatmap)
boxes = draw_labeled_bboxes(labels)
return boxes, heatmap
In [20]:
# create weighted heatmap
def find_weighted_heat(img, frames):
heat = np.zeros_like(img[:,:,0]).astype(np.float)
for i in range(len(frames)):
heat = add_heat(heat,frames[i], 1)
heat_max = max(heat.flatten())
thresh = (heat_max+1)//2
heat = apply_threshold(heat,thresh)
heatmap = np.clip(heat, 0, 255)
labels = label(heatmap)
boxes = draw_labeled_bboxes(labels)
return boxes, heatmap
In [21]:
# video processing pipeline
frames = []
count = 0
def video_pipeline(img, vis=False):
global frames
global count
box_list = []
if count % 2 == 0:
boxes = find_cars(img, 'YCrCb', 390, 662, 3, clf234, X_scaler234, orient=11, pix_per_cell=16, cell_per_block=2)
box_list = box_list + boxes
boxes = find_cars(img, 'YCrCb', 390, 662, 2.5, clf234, X_scaler234, orient=11, pix_per_cell=16, cell_per_block=2)
box_list = box_list + boxes
boxes = find_cars(img, 'YCrCb', 390, 582, 2, clf234, X_scaler234, orient=11, pix_per_cell=16, cell_per_block=2)
box_list = box_list + boxes
boxes = find_cars(img, 'YCrCb', 390, 582, 1.5, clf234, X_scaler234, orient=11, pix_per_cell=16, cell_per_block=2)
box_list = box_list + boxes
boxes = find_cars(img, 'YCrCb', 390, 486, 1, clf234, X_scaler234, orient=11, pix_per_cell=16, cell_per_block=2)
box_list = box_list + boxes
if len(frames) > 0 and len(box_list) == 0:
frames.remove(frames[0])
count = count + 1
out_img = np.copy(img)
#for box in box_list:
# cv2.rectangle(out_img,box[0],box[1],(0,255,0),6)
#
#boxes, heatmap = find_heat(img, box_list, 1)
#for box in boxes:
# cv2.rectangle(out_img,box[0],box[1],(0,0,255),6)
if len(box_list) > 0:
frames.append(box_list)
if(len(frames) > 10):
frames.remove(frames[0])
#rects = []
#for frame in frames:
# rects = rects + frame
#for box in rects:
# cv2.rectangle(out_img,box[0],box[1],(0,255,0),6)
#for box in box_list:
# cv2.rectangle(out_img,box[0],box[1],(255,255,0),6)
#n = int(math.sqrt(len(rects)))
#boxes, heatmap = find_heat(img, rects, n)
boxes, heatmap = find_weighted_heat(img, frames)
for box in boxes:
cv2.rectangle(out_img,box[0],box[1],(0,0,255),6)
if vis == True:
return out_img, heatmap
return out_img
In [22]:
# test video_pipeline and display heatmap
def try_test_images():
global frames
global count
fg, ax = plt.subplots(3, 4, figsize=(40, 18))
ax = ax.ravel()
i = 0
files = glob.glob('test_images/*.jpg')
for file in files:
# read image from file
img = read_image(file)
while(len(frames) > 0):
frames.remove(frames[0])
count = 0
out_img, heatmap = video_pipeline(img, True)
# display original image
ax[i].axis('off')
ax[i].imshow(out_img)
ax[i].set_title(file, fontsize=24)
# display transformed image
ax[i+1].axis('off')
ax[i+1].imshow(heatmap, cmap='hot')
ax[i+1].set_title('heatmap', fontsize=24)
i = i + 2
try_test_images()
In [23]:
while(len(frames) > 0):
frames.remove(frames[0])
count = 0
from moviepy.editor import VideoFileClip
from IPython.display import HTML
challenge_output = 'output_video.mp4'
## To speed up the testing process you may want to try your pipeline on a shorter subclip of the video
## To do so add .subclip(start_second,end_second) to the end of the line below
## Where start_second and end_second are integer values representing the start and end of the subclip
## You may also uncomment the following line for a subclip of the first 5 seconds
##clip3 = VideoFileClip('test_videos/challenge.mp4').subclip(0,5)
#clip3 = VideoFileClip('project_video.mp4').subclip(15,35)
clip3 = VideoFileClip('project_video.mp4')
challenge_clip = clip3.fl_image(video_pipeline)
%time challenge_clip.write_videofile(challenge_output, audio=False)
In [24]:
HTML("""
<video width="960" height="540" controls>
<source src="{0}">
</video>
""".format(challenge_output))
Out[24]: